Photo enhancement using HDR structure

For many photographers, HDR images are not a preferred option for the final piece, but HDR can be used in post-processing for intermediate stages

Many photographers dislike High Dynamic Range images, especially when overused.  Although I sympathise with this view, structure control (also known as micro-contrast) is one HDR feature that every photo retoucher should know about.

Local contrast is called ‘structure’ in HDR Efex Pro and ‘micro contrast’ in Photomatix.

Structure vs. contrast.

All RAW images have low initial contrast and increasing it is the most common retouching fix. Increasing contrast has unfortunate effects;

  • Loss of  detail, especially in highlights or shadows
  • Loss of  colour accuracy.

A fix is to use HDR structure. Structure gives better defination between pixels, enhancing textures, and returning some of the detail lost via contrast tuning. Used instead of contrast, structre can boost the local detail of an image considerably. The process can be applied to non-HDR images.

Step 1: Create a single exposure tonal HDR image with high structure/micro-contrast.

I use HDR Efex Pro/Lightroom in this step, but you can just as easily use Photomatix or Photoshop (especially Photoshop CS5).

Starting with a RAW image, export to HDR Efex Pro. Set the tone compression to -100, the saturation to -100 (giving you a black and white image).

Original RAW image
Original RAW image

Set the structure to 100 and the HDR method strength to about 50%. Select a  HDR method that gives a detailed high contrast result. Save the image.

You now have two images: the original RAW and the B&W single exposure HDR image that you just created from it.

Step 2: Merge the original and HDR tonal image in Photoshop.

Export/Load the two images into Photoshop.

Look at the original and HDR image example below.  The HDR image has enhanced details. Cracks, smudges and textures are more visible in the HDR image. By merging the two images, we will add these details in the colour image.

Comparing the original and HDR images
Comparing the original and HDR images

Add the HDR image as a new layer above the colour image. Set Layer type to Luminosity, and Opacity to 20-30%.  The image below shows the merged version (top) vs the original (bottom).  The texturing and detail is greater on the merged version: there are details that are not evident at all in the original!  The additional detail also looks entirely natural. Best of all, we kept the colours accurate despite adding new detail.

The merged image will occasionally look brighter. Fix this via Image > Adjustments > Exposure… and change gamma (make darker by 0.1-0.2 by moving the slider to the right).

Merging the original and HDR image
Merging the original and HDR image

 

Examples

Tokina 11-16 (11mm, ISO400, 1/25s at f10, hand held, no flash)
Tokina 11-16 (11mm, ISO400, 1/25s at f10, hand held, no flash)

I’m unhappy with this image. The glow effect  gives a nice ambience but  also loses detail in the foreground trees. We will regain this detail without compromising colour and ambience.

HDR map
HDR map

This is the HDR tonal map, created using the process described above. We can use the  texture information in the map to add detail into the colour image.

Enhanced merged image
Enhanced merged image

Here’s the merged image. I’ve used a layer mask to add detail only to specific areas (the foreground subjects).

Larger image here (opens in new browser/tab, 1.2Mb).

Conclusion

HDR images can be mixed with the original image to produce effects that are not easily achieveable by other means. We saw how structure/micro contrast processing can be used to enhance texture and add detail to a heavily processed image.

For many photographers, HDR images are not a preferred option for the final piece, but HDR can be used in post-processing for intermediate stages (e.g. opening up shadows, adding texture, softening highlight burnouts, etc).

Notes

  1. When creating the Black and white HDR image, you must always use the original RAW image. If you use an image processed in any way (including sharpening), you may not get acceptable results. The colour image can of course be processed (and usually is, and the reason you go through the process above at all  is because the processing has reduced local contrast and detail).
  2. The black and white image is actually a luminosity map.  It follows that this map will be better if it has a ‘high luminosity range’, or to put it into HDR terms, if it is high dynamic range. Thus, creating the luminosity map with a number of exposures (i.e. a true HDR image) will give a stronger effect. I don’t do this in the tutorial because I use the technique to to increase low level texture definition in a single image rather than add ‘toned down HDR’.
Advertisements